Robert got me looking at various Jakarta stats in a recent email of his.
Gave me something to do while I waited for a plane at 5am this morning :)
So, here's a dump of stats. A committer is defined as somebody with svn
access.
316 committers in Jakarta.
107 on PMC, 209 not on PMC.
8349 commits in 2005 from 121 committers.
(5768 from 90 committers in 2004, 30 of whom did not commit in 2005)
30 committers only have access to jakarta-pmc or jakarta-site.
Quick report of # people committing to N subprojects; not including
site/pmc and merging commons and commons-sandbox into one.
People - Components
30 - 0
176 - 1
74 - 2
18 - 3
12 - 4
3 - 5
2 - 6
1 - 9
So 30 people are not really committers, 176 only commit on one subproject
etc. Bear in mind that you have to apply a filter of 2/3rds to see the
active ones. So assuming a perfect balance, only 2 of the most spread 6
committers are actually active.
Next up. Cross-community activity. Ignoring the 206 who do not cross a
community, and the 6 who are all over the place (and probably doing infra
things, project setp), the top ten combinations are:
32 - turbine jcs
9 - commons-sandbox turbine jcs
9 - commons taglibs
7 - commons slide
4 - commons turbine jcs velocity
3 - commons-sandbox taglibs
3 - commons httpcomponents
3 - commons turbine jcs
2 - commons tapestry turbine jcs
2 - bcel commons-sandbox
Turbine/JCS is a misnomer, we copied the Turbine SVN over when setting JCS
up. Commons makes up the rest.
Ignoring Turbine/JCS, and ignoring Commons as a whole, what cross
community is there committer wise. The entire list is:
5 - turbine jcs velocity
3 - hivemind tapestry
2 - turbine jcs taglibs
2 - poi slide
2 - tapestry turbine jcs
1 - slide velocity
1 - jmeter turbine jcs
1 - cactus slide
1 - cactus turbine jcs
1 - poi tapestry
1 - slide taglibs
1 - ecs oro regexp taglibs
1 - cactus taglibs
1 - bcel taglibs
1 - hivemind slide tapestry
1 - bsf poi regexp taglibs
Some are obvious, some are because taglibs is commons-like in its
community (I think), some are unexpected.
----
The major reason for looking at this was to have scripts to point out how
many committers are not on the pmc. 2/3rds of Jakarta. Of course, 2/3rds
of Jakarta are inactive too. So back to the data.
Comparing the active committers of 2005, against the PMC list, we get two
interesting factoids:
1) Inactive PMC members : 39
2) Active non-PMC committers : 53
To finish it off:
3) Active PMC members : 68
===
Next up I guess. Building svn logs for each project instead of Jakarta as
a whole, so we can see where activity is, and whether we have oversight
problems.
Hen
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
